st: RE: panel data

["Lachenbruch, Peter" <Peter.Lachenbruch@oregonstate.edu>]

As Nick suggested, use multiple imputation.  I'm auditing a Bayesian class and an interesting idea came up and using winBUGS allows you to specify a model for the missing observations and estimate them.  Check out with your friendly local Bayesian and see what can be done.  You'll need to learn winBUGS.  I understand there is a program that allows you to call winBUGS from Stata (Stata2BUGS?)
TONY

________________________________________
From: owner-statalist@hsphsun2.harvard.edu [owner-statalist@hsphsun2.harvard.edu] On Behalf Of Deepti Garg [deeptigarg78@yahoo.com]
Sent: Friday, January 27, 2012 5:50 AM
To: statalist@hsphsun2.harvard.edu
Subject: st: panel data

Hi,


I have around 170 firms and around 16 variables in a panel data with firms having data for at the most 5 years. However,  I dont have data for all the firms for all the variables. So, when I run regression command, stata only picks the data which has the data for all the variables and my results are inappropriate.

However, when I run the regression with each independent variable separately, I get the desired results. Is it valid to run separate regressions for all the variables and analyse the results. Or, is there any way where I can force stata to consider records which do not have all the data and take a best case scenario?

Any help is appreciated.

Thanks!
Deepti

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/